1. Field of the Invention
This invention relates to a power converter device for interconnecting alternating current systems, and more particularly to a power converter device wherein the alternating current windings of a plurality of transformer units are connected in series, and are respectively connected to a plurality of self-commutated voltage type converter units.
2. Discussion of Background
FIG. 6 shows a schematic block diagram of a conventional widely used power converter device for interconnecting alternating current systems. Reference numerals 11 and 21 designate three-phase self-commutated voltage type inverter units (hereinbelow called inverter units); 12, 22 designate system interconnecting reactors; 13, 23 designate isolating transformer units; 14, 24 designate windings of transformer units 13, 23; 15, 25 likewise designate a.c. windings.
The a.c. windings 15 and 25 of transformer units 13, 23 are connected in series in each phase. Their outputs are connected to a.c. system 32 through a.c. switches 31. In contrast, both the d.c. sides of inverter units 11 and 21 have a common d.c. power source 33. A capacitor 34 is connected to the d.c. circuit. The a.c. windings 15 and 25 of transformer units 13 and 23 form what is called a zig-zag connection. This is a widely used method of connection in order to obtain the advantage that, when inverter units 11 and 21 are operated with a mutual phase difference of 30.degree., no harmonic components other than the (12p.+-.1)-th order harmonics (p=1, 2 ...) are contained in the resultant output of transformer units 13 and 23.
Operation of the conventional power converter device for system interconnection shown in FIG. 6 is initiated as follows. Inverter units 11, 21 are started in operation by a signal from a control circuit, not shown. The rise of the output voltage of inverter units 11, 21 from zero is made gradual, to avoid excitation rush current in transformer units 13, 23. A.C. switches 31 are closed when the amplitude and phase of the resultant voltage of a.c. windings 15, 25 of transformer units 13, 23 coincide with the amplitude and phase of the voltage of system 32. This is called synchronous making. In case of malfunction or interruption, a.c. switches 31 are opened simultaneously with cessation of operation of inverter units 11, 21. The time between stopping of inverter units 11, 21 and opening of a.c. switches 31 is less than 0.1 second, to avoid the problem to be discussed below.
Self-commutated inverters for system interconnection have recently begun to be used in systems such as solar generators or fuel cells. System interconnection based on self-commutated inverters, in the case where the a.c. system is weak, are of superior stability compared with system interconnection based on externally commutated inverters. The reason for this is that, in the case of externally commutated inverters commutation of the thyristors depends on the voltage of the a.c. system, so commutation of the thyristors may be prevented by disturbance of the system voltage, i.e. so-called commutation failure occurs. To deal with this, in the case of a self-commutated inverter, commutation of the thyristor is performed by a commutation circuit within the inverter, so disturbance of the system voltage does not immediately result in commutation failure. However, in the aforementioned operation control system, operation of inverter units 11, 21 may be temporarily cut off by generation of a.c. overcurrent if for example there is an instantaneous voltage fluctuation of the a.c. system. In this case, it is difficult to restart the operation immediately when the voltage is reset. The reason for this is that a.c. switches 31 must open whenever a.c. overcurrent occurs. This gives rise to the problem that: (a) once stoppage has occurred, about five seconds is required from recommencement of operation until synchronous making, and (b) the life of a.c. switches 31 is adversely affected by frequent switching of a.c. switches 31.
If it is assumed that an operating scheme is chosen according to which closure of a.c. switches 31 is followed by operation of the inverter, it might be thought that, for the aforementioned overcurrent protection, it would be sufficient simple to stop the operation of the inverter, i.e. simply to stop the on/off operation of the gate of the GTO. This ought to have the effect of stopping a.c. switches 31 from opening. However, it has been found that there are two problems in doing this. One is the d.c. overvoltage generated in the transient period when the connection of the transformers to the system is closed, and the other is the d.c. overvoltage in the steady state.
First discussed is the d.c. overvoltage in the transient period. Since, as shown in FIG. 6, a.c. windings 15 and 25 are connected in series, if, at the instant when a.c. switches 31 are closed, the a.c. voltage is not applied to the two transformers 13 and 23 equally, the result is the appearance of an alternating current having a high peak, quite different from a sine wave. The first cause of this is that the initial magnetization state of the two transformers is not the same. The second cause is that there is high impedance for high frequencies of harmonic order other than 12p.+-.1(p=1, 2 ...), because of the zig-zag connection, so the harmonic exciting current components needed to induce a sine wave voltage cannot flow. That is, the excitation characteristic of the iron core is not totally linear, but rather is a non-linear curve having hysteresis. Since, in order to create a sine wave voltage, high frequencies must be contained in the exciting current, if there is some restriction that prevents these high frequencies from flowing, the induced voltage will not be a sine wave.
In experiments, on transition, the peak value of the voltage induced in d.c. windings 14 and 24 of transformer units 13 and 23 was about 2.9 times the root mean square value of the sine wave. The d.c. voltage of the capacitor 34 is therefore charged up to this value.
The second problem concerns d.c. overvoltage in the steady state. As described above, the induced voltage is not a simple sine wave, but contains harmonic components. It has been experimentally determined that a peak value of about 2.4 times the root mean square value of the sine wave appears. The d.c. voltage of the capacitor 34 was charged up to this value. FIG. 8 shows measured waveforms obtained by experiment as described above.
In the above description, it is assumed that inverter units 11 and 21 are employed. However, this invention is applicable not merely to conversion from d.c. to a.c., but also to reactive power compensator devices, or rectifiers whose power factor can be regulated, so hereinbelow, numerals 11 and 21 refer generally to self-commutated voltage type converters, and the whole system will be taken as being a power converter device.